Below you will find the PDF file describing the final project.

Due date: 18.1.2020

You are expected to provide not only the Monte Carlo estimate but also all supporting analysis that validates your MCMC algorithm.

Q&A

1. Question: I don't understand the purpose of the f array.

Answer: The f array gives you the frequencies of data points in a given block of the parameter space. And you use this in calculating your acceptance probability. So that probability of acceptance is equal to min(1,Fnew/Fold).

2. Question: I am getting the error message: Unable to perform assignment because the left and right sides have a different number of elements.

Error in updateFSA (line 13)

       qc(j) = executeEvents(FSA{j},s(i),qc(j));


Error in NetworkSimRun (line 99)

           qSupnew = updateFSA(FSAsup,enew,qSupc);


3. Question: How do we compute the error?

Answer: The data set alone will not give you an accurate estimate. You must generate more samples. Then you have three methods by which you can compute the error: z-score, t-score, boostrap. Which ever method you use, justify your answers. Recall the z-score assumes the sample standard deviation is the same as the actual standard deviation. The t-score and the z-score both depend on the central limit theorem.

Please note that in problem 2, when comparing abnormal time estimated by MCMC to the abnormal time estimated from your simple Markov Chain model should only consider abnormal voltage at the second node (not the reactive power).

Assigned datasets (from the list below) will be mailed to you upon request (be email):

1. DataSet100-01.mat
2. DataSet100-02.mat
3. DataSet100-03.mat
4. DataSet100-04.mat
5. DataSet100-05.mat
6. DataSet100-06.mat
7. DataSet100-07.mat
8. DataSet100-08.mat
9. DataSet100-09.mat
10. DataSet100-10.mat
11. DataSet100-11.mat
12. DataSet100-12.mat
13. DataSet100-13.mat
14. DataSet100-14.mat
15. DataSet100-15.mat